Inductive reasoning and deductive reasoning are similar in that they are two logical approaches to exploring data and to drawing conclusions. They differ in that they approach logical, research, philosophical, psychological or other problems from differing perspectives, one moving from the specific to the general and the other moving from the general to the specific. Induction begins with specific ideas, details, data or observations and ends with general principles, conclusions, concepts, rules, hypotheses or theories, while, in the converse, deduction begins with general overall principles, rules, concepts, conclusions, hypotheses or theories and ends with specific data, ideas, details or observations (these are then used to narrow down the hypotheses and clarify the theories they support or contradict).
While the concepts of inductive and deductive logical reasoning are the same regardless of the labels used to identify and refer to them, there are differing ways of referring to the processes of inductive (specific-to-general) and deductive (general-to-specific) reasoning. Sometimes, the differences in reference are a result of the field in which induction and deduction are being applied to research, logic or experimentation. For example, regarding inductive specific-to-general reasoning, in the field of astronomy, the reference may be to observation and hypothesis while in philosophy or in the study of logic the reference may be to premise and conclusion. Whereas in various fields of research, inductive reasoning may be referred to as bottom-up reasoning as it is based on individual ideas and data gathered from experiment participants or results, conversely, deductive (general-to-specific) reasoning in research may be referred to as top-down reasoning as it is founded in theory and hypothesis, then supported or contradicted by collected data. To refine inductive and deductive reasoning in research further, inductive reasoning methods may be called qualitative analysis while deductive reasoning methods may be called quantitative analysis (Soiferman, "Research Approaches").
Inductive and deductive reasoning are the same in that both are built upon the specific and the general, and both are applied across varying fields of study and investigation, and both may be used to explore the same questions. Yet they are different in that they represent different methodologies, and they have different starting points and end points within the specific and the general, and they have different ways of and focuses on examining truths across wide disciplines and fields of study.
Another important difference that comes from having unlike starting and end points (i.e., either specific or general starting points or end points) relates to the certainty or probability of reaching true conclusions. Because it moves from the specific to the general, inductive reasoning can result in only probabilities of truth: general conclusions cannot be certain truths when starting data or premises, starting specifics, are themselves not proven. Because the other moves from the general to the specific, deductive reasoning can result in certainty (varying degrees of certainty): specific situations can be understood with certainty when the starting principles of hypothesis and theory are true.
Note: It's not easy to find clear, easily understandable explanations of inductive and deductive reasoning online. The explanatory language is often vague or confusing, with terms that are interchanged without clear indication as to why. One of the clearest is provided by ERIC Institution of Education Science, eric.ed.gov (under the U.S. Department of Education), in an article called "Inductive and Deductive Research Approaches" by L. Karen Soiferman of the University of Manitoba, from which some of this answer is drawn. Another useful explanation, though not without its confusingly interchangeable words and confusing sentences, is provided by Norman Herr, Ph.D. of the California State University of Northridge science department. One of the difficulties in understanding these logical reasoning concepts is that while the component parts are the same (e.g., premise-conclusion, hypothesis-observable data), the use and end results are very different. As Norman Herr states it:
Scientists use inductive reasoning to formulate hypothesis [sic] and theories, and [use] deductive reasoning when applying them [i.e., hypotheses and theories] to specific situations. ("Deductive Reasoning")